[REBOL] Re: Rebol Fluid
From: d4marcus:dtek:chalmers:se at: 13-Dec-2000 18:51
I figured I shouldn't let this pass without a comment...
On Sun, 10 Dec 2000, Petr Krenzelok wrote:
> Although KOSH project seems to be still alive, I think we will meet with
> most of its principles via upcoming
> Amiga Digital Environment, as it is designed by Fleecy Moss (currently Amiga
> Inc.) .... It will be imo cool.
No doubt, I'm looking forward to it too. AmigaDE is possibly the only real
advancement in mainstream OS technology since BeOS. Well, maybe you could
count QNX too, but that isn't exactly mainstream, and besides, parts of
its design is older than BeOS. But the Elate/Amiga stuff seems promising.
Still, it seems to carry some old baggage with it, for example the concept
of a filesystem. This may not seem important, but it does impose some
restrictions on how object oriented the environment can be. Which in its
turn affects user experience. So, even if AmigaDE means a revolution for
both system-level development and user experience, it will not be a total
revolution as long as it comes with the filesystem legacy. Hopefully, they
will be developing a system object database sometime, I'm quite sure this
is Fleecy's intention anyway. But you don't always get what you want.
> As for KOSH principles - we are near. Just look at current rebol
> architecture, look how VID is created, how it works etc.
> IMO Rebol is the nearest thingy to KOSH principles nowadays .....
Very nice, but not really close to KOSH principles. Someone started the
question of a database written in Rebol, and chaz mentioned he'd use it to
store face components. I think this is perfectly reasonable idea, and it's
something I would appriciate as well. By all means, go ahead and create
If you could store REBOL code and data objects like that, you can achieve
lots of benifits. "There's no end to the possibilities". :-) One thing
you could do (or so I think) is to create a dialect, that each time an
unknown word is called looks it up in the database. You get your own Rebol
repository, a sort of library containing all your Rebol code and data,
even non-Rebol data (but in base64 form).
You wouldn't need to use files at all, except when you install ("import")
something new. This of course includes editing existing code, unless you
had a decent Rebol editor, or could convince an existing editor to talk to
the database. And whenever you need something from the database in a file,
you call the "export" function.
Still quite a bit from the object sea idea we have in KOSH. But if you
could boot the computer directly into Rebol, with or without View, run
some simple filesystem or DBMS underneath, then have drivers talk to the
hardware, essentially have your entire working environment within Rebol -
then you would be close.
But I don't yet see such a thing, in Rebol or anything else. That's why
we're continuing KOSH.
If you find that life spits on you
calm down and pretend it's raining